Can AI have taste? That was the hot topic at Replit's NYC vibe-coding conference.

The question of whether artificial intelligence can develop "taste"—a term increasingly used in tech circles to describe intuitive, human-like judgment in design, aesthetics, and creativity—dominated discussions at Replit’s recent Vibecon in New York. While the word might sound whimsical, it cuts to the heart of a critical tension shaping the future of AI: the gap between capability and cultural resonance. As generative tools grow more sophisticated, their outputs risk remaining sterile, derivative, or outright bizarre unless they can internalize the subjective, often unspoken rules that govern human aesthetic preferences. Vibecon’s focus on this issue signals a pivot from mere functionality to something far more elusive: cultural fit. This isn’t just an academic debate. The AI industry has long prioritized scale and speed, churning out content, code, and art at unprecedented volumes. Yet the backlash against formulaic, soulless outputs—whether in AI-generated music, bland corporate branding, or even bug-ridden code—has grown louder. The concept of "taste" here serves as a proxy for a deeper challenge: teaching machines to understand context, subtext, and emotional nuance. Unlike traditional programming, which relies on rigid logic, taste requires an intuitive grasp of what feels right, culturally appropriate, or simply *good*—qualities that elude even the most advanced models. What comes next is unclear. Will AI systems ever truly develop taste, or will they remain sophisticated mimics, forever one step behind human intuition? Some argue that the answer lies in better training data—curating datasets that reflect diverse cultural expressions. Others suggest hybrid models, blending algorithmic generation with human oversight. But the conversation at Vibecon hints at a more radical possibility: that taste isn’t something AI can be programmed to have, but rather something it must co-create with humans over time. If that’s the case, the next frontier of AI development may not be about smarter models, but about richer, more collaborative relationships between humans and machines. The implications stretch beyond tech, touching industries from entertainment to design. If AI can’t grasp taste, its role may remain confined to the mundane—filling gaps rather than innovating. The stakes are high, and the answer could redefine whether AI becomes a tool of cultural enrichment or just another factory of mediocrity.